DESCRIPTION: The investigators endorse a constructivist perspective on children's early syntactic development, according to which children acquire linguistic competence in the particular language they are learning only gradually, beginning with more concrete linguistic structures based on particular words and morphemes then building up to more abstract and productive structures based on various types of linguistic categories, schemas, and constructions. In contrast, nativist models, such as Pinker's, argue that basic syntactic categories are innate and so children's basic task is to discover how these categories are instantiated in the particular languages they are learning. Constructivist models suggest that children actually construct the basic syntactic categories of the particular languages they are learning from concrete exemplars of adult speech. The Tomasello-Brooks view contrasts most radically with the views of more nativistically oriented linguists who believe that children's linguistic competence is generated from the beginning by abstract syntactic principles and parameters that constitute a coherent and integrated formal grammar. But it also contrasts with most constructivists theories that focus on grammatical categories in isolation from grammatical constructions as the basic phenomena to be explained. Thus, when a child says something like "I kick ball" is their underlying competence best characterized as consisting of an abstract schema such as NP-V-NP or is it best characterized as consisting of something more concrete such as ANIMATE-kick-INANIMATE? Tomasello's earlier work (1992) supports the verb-specific-schemas such as ANIMATE-kick-INANIMATE. From a construction grammar point of view the two key issues in the study of first language acquisition are: 1) when and how do young children begin to construct relatively abstract syntactic constructions?; and 2) when do they begin to constrain these abstract constructions to the conceptual situations and lexical items with which they are conventionally used in adult language? The proposed studies are designed to explore these two issues: how children acquire and constrain their earliest syntactic constructions that have at least some abstractness (i.e., that apply across a number of different verbs). The first step involves documenting the range of constructions that young children use in their natural conversational interactions with others. A descriptive analysis of ten data sets from the CHILDES collection will be used. Three- and four-year olds produce a fair number of non-canonical utterances that seem to be overgeneralization of one or another construction, for example, "Don't giggle me." The second step is to investigate experimentally how children learn and then constrain some of the most basic and abstract verb-argument constructions of English.